U.S. patent number 11,023,744 [Application Number 15/944,192] was granted by the patent office on 2021-06-01 for road parameter calculator.
This patent grant is currently assigned to DENSO CORPORATION. The grantee listed for this patent is DENSO CORPORATION. Invention is credited to Taiki Kawano, Naoki Kawasaki, Shunya Kumano, Shunsuke Suzuki.
United States Patent |
11,023,744 |
Kawano , et al. |
June 1, 2021 |
Road parameter calculator
Abstract
A road parameter calculator is provided which is equipped with
an edge-point extracting unit, a road parameter calculating unit, a
gradient detecting unit, and a modeling unit. The image acquiring
unit. The edge-point extracting unit extracts edge points from an
image of a frontal view of a vehicle. The parameter calculating
unit calculates a road parameter using the edge points through a
Kalman filter. The gradient detecting unit detects a change in
gradient of the road in front of the vehicle. The modeling unit is
responsive to a change in gradient to make a model as extending
more straight than when the change in gradient is not detected.
This minimizes adverse effects of the change in gradient of the
road on the calculation of the road parameter.
Inventors: |
Kawano; Taiki (Aichi-pref.,
JP), Kawasaki; Naoki (Aichi-pref., JP),
Suzuki; Shunsuke (Aichi-pref., JP), Kumano;
Shunya (Aichi-pref., JP) |
Applicant: |
Name |
City |
State |
Country |
Type |
DENSO CORPORATION |
Kariya |
N/A |
JP |
|
|
Assignee: |
DENSO CORPORATION (Kariya,
JP)
|
Family
ID: |
1000005590530 |
Appl.
No.: |
15/944,192 |
Filed: |
April 3, 2018 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20180293447 A1 |
Oct 11, 2018 |
|
Foreign Application Priority Data
|
|
|
|
|
Apr 5, 2017 [JP] |
|
|
JP2017-075015 |
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06T
7/277 (20170101); G06K 9/4604 (20130101); G06T
7/12 (20170101); G06K 9/00798 (20130101); G06T
7/73 (20170101); B60W 2552/20 (20200201); B60W
2552/05 (20200201); G06T 2207/20061 (20130101); G06T
2207/30256 (20130101); G06T 2207/20076 (20130101) |
Current International
Class: |
G06K
9/00 (20060101); G06T 7/12 (20170101); G06T
7/277 (20170101); G06T 7/73 (20170101); G06K
9/46 (20060101) |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Safaipour; Bobbak
Attorney, Agent or Firm: Maschoff Brennan
Claims
What is claimed is:
1. A road parameter calculator comprising: an image acquiring unit
which is configured to acquire an image of a forward view of a
vehicle; an edge-point extracting unit which is configured to
extract edge points from the image derived by the image acquiring
unit; a road parameter calculating unit which is configured to
calculate a road parameter through a Kalman filter using the edge
points derived by the edge-point extracting unit, the road
parameter representing a configuration of a road in front of the
vehicle; a gradient detecting unit which is configured to detect a
change in gradient of the road in front of the vehicle; and a
modeling unit which is configured to produce a model used in the
Kalman filter, when the gradient detecting unit detects the change
in gradient of the road, the modeling unit working to produce a
model of the road in which a curvature of the road is reduced than
when the change in gradient of the road is not detected.
2. A road parameter calculator as set forth in claim 1, wherein
when the gradient detecting unit detects the change in gradient,
and only one of a right lane line and a left lane line is detected,
the modeling unit works to produce the model designed to model the
road in which the curvature of the road is reduced further than
that when the gradient detecting unit detects the change in
gradient, and the right and left lane lines are both detected, the
right and left lane lines being lines on the image which define a
road on which the vehicle is currently located.
3. A road parameter calculator as set forth in claim 1, further
comprising a response setting unit which sets a response rate of
the Kalman filter, and wherein when the gradient detecting unit
detects the change in gradient, the response setting unit sets the
response rate of the Kalman filter to be lower than that when the
gradient detecting unit detects no change in gradient.
4. A road parameter calculator as set forth in claim 3, further
comprising a degree-of-gradient calculating unit which is
configured to calculate a degree of the change in gradient, and
wherein the response setting unit lowers the response rate with an
increase in degree of the change in gradient, as calculated by the
degree-of-gradient calculating unit.
5. A road parameter calculator as set forth in claim 1, wherein
when a first vanishing point remains above or below a second
vanishing point, said gradient detecting unit determines that there
is the change in gradient, the first vanishing point being an
intersection of a right lane line and a left lane line, the right
and left lane lines being lines on the image which define a lane on
which the vehicle now exist, the second vanishing point being an
intersection of the right and left lane lines when the vehicle is
moving on a flat and horizontal road surface.
6. A road parameter calculator as set forth in claim 1, wherein the
gradient detecting unit detects the change in gradient using a
sensor which is capable of measuring a configuration of a surface
of the road.
7. A road parameter calculator as set forth in claim 1, wherein the
gradient detecting unit detects the change in gradient using a
change in positional relation between a preceding vehicle and a
lane line in a width-wise direction of the vehicle or a positional
relation between a locus of the preceding vehicle and the lane
line.
8. A road parameter calculator as set forth in claim 1, wherein the
gradient detecting unit detects the change in gradient using a
combination of curvatures of a right and a left lane line which are
lines on the image defining a road on which the vehicle is
currently located.
9. A road parameter calculator as set forth in claim 1, wherein the
gradient detecting unit detects the change in gradient using
information about gradients of roads recorded in a map.
10. A road parameter calculator comprising: an image acquiring unit
which is configured to acquire an image of a forward view of a
vehicle; an edge-point extracting unit which is configured to
extract edge points from the image derived by the image acquiring
unit; a road parameter calculating unit which is configured to
calculate a road parameter through a Kalman filter using the edge
points derived by the edge-point extracting unit, the road
parameter representing a configuration of a road in front of the
vehicle; a gradient detecting unit which is configured to detect a
change in gradient of the road in front of the vehicle; and an
edge-point limiting unit which is configured to define a range in
which the edge points are used by the road parameter calculating
unit to lie closer to the vehicle when the gradient detecting unit
detects the change in gradient than when the gradient detecting
unit does not detect the change in gradient.
11. A road parameter calculator as set forth in claim 10, wherein
when a first vanishing point remains above or below a second
vanishing point, said gradient detecting unit determines that there
is the change in gradient, the first vanishing point being an
intersection of a right lane line and a left lane line, the right
and left lane lines being lines on the image which define a lane on
which the vehicle now exist, the second vanishing point being an
intersection of the right and left lane lines when the vehicle is
moving on a flat and horizontal road surface.
12. A road parameter calculator as set forth in claim 10, wherein
the gradient detecting unit detects the change in gradient using a
sensor which is capable of measuring a configuration of a surface
of the road.
13. A road parameter calculator as set forth in claim 10, wherein
the gradient detecting unit detects the change in gradient using a
change in positional relation between a preceding vehicle and a
lane line in a width-wise direction of the vehicle or a positional
relation between a locus of the preceding vehicle and the lane
line.
14. A road parameter calculator as set forth in claim 10, wherein
the gradient detecting unit detects the change in gradient using a
combination of curvatures of a right and a left lane line which are
lines on the image defining a road on which the vehicle now
exists.
15. A road parameter calculator as set forth in claim 10, wherein
the gradient detecting unit detects the change in gradient using
information about gradients of roads recorded in a map.
16. A road parameter calculator comprising: an image acquiring unit
which is configured to acquire an image of a forward view of a
vehicle; an edge-point extracting unit which is configured to
extract edge points from the image derived by the image acquiring
unit; a road parameter calculating unit which is configured to
calculate a road parameter through a Kalman filter using the edge
points derived by the edge-point extracting unit, the road
parameter representing a configuration of a road in front of the
vehicle; a gradient detecting unit which is configured to detect a
change in gradient of the road in front of the vehicle; and an
edge-point limiting unit which is configured to define a right
range and a left range in which the edge points on a right lane
line and a left lane line are used by the road parameter
calculating unit, the right and left lane lines being lines on the
image which define a road on which the vehicle now exists, when the
gradient detecting unit detects the change in gradient, the
edge-point limiting unit setting the right and left ranges to be
equal to each other.
17. A road parameter calculator as set forth in claim 16, wherein
when a first vanishing point is continuously kept above or below a
second vanishing point, said gradient detecting unit determines
that there is the change in gradient, the first vanishing point
being an intersection of a right lane line and a left lane line,
the right and left lane lines being lines on the image which define
a lane on which the vehicle now exist, the second vanishing point
being an intersection of the right and left lane lines when the
vehicle is moving on a flat and horizontal road surface.
18. A road parameter calculator as set forth in claim 16, wherein
the gradient detecting unit detects the change in gradient using a
sensor which is capable of measuring a configuration of a surface
of the road.
19. A road parameter calculator as set forth in claim 16, wherein
the gradient detecting unit detects the change in gradient using a
change in positional relation between a preceding vehicle and lane
line in a width-wise direction of the vehicle or a positional
relation between a locus of the preceding vehicle and the lane
line.
20. A road parameter calculator as set forth in claim 16, wherein
the gradient detecting unit detects the change in gradient using a
combination of curvatures of a right and a left lane line which are
lines on the image defining a road on which the vehicle is
currently located.
21. A road parameter calculator as set forth in claim 16, wherein
the gradient detecting unit detects the change in gradient using
information about gradients of roads recorded in a map.
22. A road parameter calculator comprising: an image acquiring unit
which is configured to acquire an image of a forward view of a
vehicle; an edge-point extracting unit which is configured to
extract edge points from the image derived by the image acquiring
unit; a road parameter calculating unit which is configured to
calculate a road parameter through a Kalman filter using the edge
points derived by the edge-point extracting unit, the road
parameter representing a configuration of a road in front of the
vehicle; a gradient detecting unit which is configured to detect a
change in gradient of the road in front of the vehicle; a lane line
producing unit which is configured to produce a lane line using the
edge points derived by the edge-point extracting unit; a likelihood
calculating unit which is configured to calculate a likelihood that
the lane line, as derived by the lane line producing unit, is a
branch line using at least one of a curvature, a yaw angle, and an
offset between the vehicle and the lane line; a branch line
determining unit which is configured to determine that the lane
line represents the branch line when the likelihood, as derived by
the likelihood calculating unit, is greater than a threshold value;
an edge-point removing unit which is configured to remove ones of
the edge points which define the lane line, as determined by the
branch line determining unit as being the branch line, from a range
in which the edge points are used by the road parameter calculating
unit; and a threshold determining unit which is configured to
increase the threshold used when the gradient detecting unit
detects the change in gradient to be greater than that when the
gradient detecting unit does not detect the change in gradient.
23. A road parameter calculator as set forth in claim 22, wherein
when a first vanishing point remains above or below a second
vanishing point, said gradient detecting unit determines that there
is the change in gradient, the first vanishing point being an
intersection of a right lane line and a left lane line, the right
and left lane lines being lines on the image which define a lane on
which the vehicle is currently located, the second vanishing point
being an intersection of the right and left lane lines when the
vehicle is moving on a flat and horizontal road surface.
24. A road parameter calculator as set forth in claim 22, wherein
the gradient detecting unit detects the change in gradient using a
sensor which is capable of measuring a configuration of a surface
of the road.
25. A road parameter calculator as set forth in claim 22, wherein
the gradient detecting unit detects the change in gradient using a
change in positional relation between a preceding vehicle and a
lane line in a width-wise direction of the vehicle or a positional
relation between a locus of the preceding vehicle and the lane
line.
26. A road parameter calculator as set forth in claim 22, wherein
the gradient detecting unit detects the change in gradient using a
combination of curvatures of a right and a left lane line which are
lines on the image defining a road on which the vehicle is
currently located.
27. A road parameter calculator as set forth in claim 22, wherein
the gradient detecting unit detects the change in gradient using
information about gradients of roads recorded in a map.
Description
CROSS REFERENCE TO RELATED DOCUMENT
The present application claims the benefit of priority of Japanese
Patent Application No. 2017-75015 filed on Apr. 5, 2017 the
disclosure of which is incorporated herein by reference.
BACKGROUND
1 Technical Field
The invention relates generally to a road parameter calculator.
2 Background Art
Japanese Patent First Publication No. 2011-28659 (corresponding to
US 2012/0185167 A1 assigned to Hitachi Automotive Systems Itd.,
disclosure of which is incorporated herein by reference) teaches a
road parameter calculator designed to capture an image of a forward
view from a vehicle using an in-vehicle camera, detect edge points
on the captured image, and calculate a road parameter using the
edge points by means of a Kalman filter.
There may be a change in gradient of a road in front of the
vehicle. This results in a difficulty in calculating the road
parameter correctly. For instance, when a lane line on the road is
actually straight, it may be determined as being a curved line due
to the gradient of the road.
SUMMARY
It is an object of this disclosure to provide a road parameter
calculator which minimizes adverse effects of a change in gradient
of a road on calculation of a road parameter.
According to one aspect of this disclosure, there is provided a
road parameter calculator which comprises: (a) an image acquiring
unit which is configured to acquire an image of a forward view of a
vehicle; (b) an edge-point extracting unit which is configured to
extract edge points from the image derived by the image acquiring
unit; (c) a road parameter calculating unit which is configured to
calculate a road parameter through a Kalman filter using the edge
points derived by the edge-point extracting unit; (d) a gradient
detecting unit which is configured to detect a change in gradient
of a road in front of the vehicle; and (f) a modeling unit which is
configured to produce a model in the Kalman filter, when the
gradient detecting unit detects the change in gradient. The
modeling unit works to produce the model which models a road
extending more straight than when the gradient change is not
detected.
The road parameter calculator is capable of minimizing adverse
effects of the change in gradient of the road on calculation of the
road parameter.
According to the second aspect of this disclosure, there is
provided a road parameter calculator which comprises: (a) an image
acquiring unit which is configured to acquire an image of a forward
view of a vehicle; (b) an edge-point extracting unit which is
configured to extract edge points from the image derived by the
image acquiring unit; (c) a road parameter calculating unit which
is configured to calculate a road parameter through a Kalman filter
using the edge points derived by the edge-point extracting unit;
(d) a gradient detecting unit which is configured to detect a
change in gradient of a road in front of the vehicle; and (e) an
edge-point limiting unit which is configured to define a range in
which the edge points are used by the road parameter calculating
unit to lie closer to the vehicle when the gradient detecting unit
detects the change in gradient than when the gradient detecting
unit does not detect the change in gradient.
The road parameter calculator in the second aspect is capable of
minimizing adverse effects of the change in gradient of the road on
calculation of the road parameter.
According to the third aspect of this disclosure, there is provided
a road parameter calculator which comprises: (a) an image acquiring
unit which is configured to acquire an image of a forward view of a
vehicle; (b) an edge-point extracting unit which is configured to
extract edge points from the image derived by the image acquiring
unit; (c) a road parameter calculating unit which is configured to
calculate a road parameter through a Kalman filter using the edge
points derived by the edge-point extracting unit; (d) a gradient
detecting unit which is configured to detect a change in gradient
of a road in front of the vehicle; and (e) an edge-point limiting
unit which is configured to define a right and a left range in
which the edge points on a right and a left lane line are used by
the road parameter calculating unit. The right and left lane lines
are lines on the image which define a road on which the vehicle is
currently located. When the gradient detecting unit detects the
change in gradient, the edge-point limiting unit sets the right and
left ranges to be equal to each other.
The road parameter calculator in the third is capable of minimizing
adverse effects of the change in gradient of the road on
calculation of the road parameter.
According to the fourth aspect of this disclosure, there is
provided a road parameter calculator which comprises: (a) an image
acquiring unit which is configured to acquire an image of a forward
view of a vehicle; (b) an edge-point extracting unit which is
configured to extract edge points from the image derived by the
image acquiring unit; (c) a road parameter calculating unit which
is configured to calculate a road parameter through a Kalman Kalman
filter using the edge points derived by the edge-point extracting
unit; (d) a gradient detecting unit which is configured to detect a
change in gradient of a road in front of the vehicle; (e) a lane
line producing unit which is configured to produce a lane line
using the edge points derived by the edge-point extracting unit;
(f) a likelihood calculating unit which is configured to calculate
a likelihood that the lane line, as derived by the lane line
producing unit, is a branch line using at least one of a curvature,
a yaw angle, and an offset of the vehicle lane line; (g) a branch
line determining unit which is configured to determine that the
lane line represents the branch line when the likelihood, as
derived by the likelihood calculating unit, is greater than a given
threshold value; (h) an edge-point removing unit which is
configured to remove ones of the edge points which define the lane
line, as determined by the branch line determining unit as being
the branch line, from a range in which the edge points are used by
the road parameter calculating unit; and (i) a threshold
determining unit which is configured to increase the threshold used
when the gradient detecting unit detects the change in gradient to
be greater than that when the gradient detecting unit does not
detect the change in gradient.
This minimizes an error in determining that the lane line is the
branch line when there is the change in gradient, thereby ensuring
the stability in calculating the road parameter correctly
regardless of the change in gradient in front of the vehicle.
BRIEF DESCRIPTION OF THE DRAWINGS
The present invention will be understood more fully from the
detailed description given hereinbelow and from the accompanying
drawings of the preferred embodiments of the invention, which,
however, should not be taken to limit the invention to the specific
embodiments but are for the purpose of explanation and
understanding only.
In the drawings:
FIG. 1 is a block diagram which illustrates a road parameter
calculator according to an embodiment;
FIG. 2 is a block diagram which illustrates a functional structure
of the road parameter calculator of FIG. 1;
FIGS. 3(a) and 3(b) are a flowchart of a program executed by a road
parameter calculator according to an embodiment.
FIG. 4 is an explanatory view which shows how to detect a change in
gradient of a road;
FIG. 5 is an explanatory view which shows how to set a range of
edge points;
FIG. 6 is an explanatory view which shows how to determine whether
there is a branch line;
FIG. 7 is an explanatory view which shows how to detect a change in
gradient of a road;
FIG. 8 is an explanatory view which shows how to detect a change in
gradient of a road;
FIG. 9 is an explanatory view which shows how to detect a change in
gradient of a road; and
FIG. 10 is an explanatory view which shows how to set a range of
edge points.
DESCRIPTION OF THE PREFERRED EMBODIMENT
An embodiment of this disclosure will be described below with
reference to the drawings.
1 Structure of Road Parameter Calculator
Referring to FIGS. 1 and 2, there is shown the road parameter
calculator 1. The road parameter calculator 1 is a vehicle-mounted
system which is installed in a vehicle such as an automobile. In
the following discussion, the vehicle equipped with the road
parameter calculator 1 will also be referred to as a
system-equipped vehicle.
The road parameter calculator 1 is made of a known microcomputer
equipped with a CPU 3 and a semiconductor memory 5 which includes,
for example, a RAM, a ROM, and a flash memory. The road parameter
calculator 1 has a variety of functions which are achieved by
executing programs, as stored in a non-transitory tangible storage
media, using the CPU 3. In this embodiment, the memory 5 is the
non-transitory tangible storage media. The programs are executed to
perform given sequences of steps. The road parameter calculator 1
may be implemented by one or more microcomputers.
The road parameter calculator 1, as illustrated in FIG. 2, includes
the image acquiring unit 7, the edge-point extracting unit 9, the
road parameter calculating unit 11, the gradient detecting unit 13,
the modeling unit 15, the response setting unit 17, the edge-point
limiting unit 19, the lane line producing unit 21, the likelihood
calculating unit 23, the branch line determining unit 25, the
edge-point removing unit 27, the threshold determining unit 29, and
the output unit 31 as units which execute the programs in the CPU 3
of the road parameter calculator 1 to achieve the functions. The
above elements which perform all or some of the functions of the
road parameter calculator 1 do not need to be realized by software,
but may be made using one or more hardware devices. For instance,
in a case where the above functions are created by hardware
electronic circuits, the electronic circuits may be made of digital
circuits equipped with many logical circuits, analog circuits, or a
combination thereof.
The system-equipped vehicle is, as illustrated in FIG. 1, equipped
with the camera 33, the surroundings sensor 35, the
quantity-of-vehicle state sensor 37, the navigation system 39, the
map information storage 41, and the driver-assistance system 43 in
addition to the road parameter calculator 1.
The camera 33 captures an image of a view in front of the
system-equipped vehicle and output it to the road parameter
calculator 1 as representing a forward view of the system-equipped
vehicle. The location and orientation of the camera 33 are always
fixed relative to the system-equipped vehicle. The surroundings
sensor 35 detects an object, such as another vehicle, a pedestrian,
or a feature (also called a landmark), around the system-equipped
vehicle. The surroundings sensor 35 is also capable of measuring
the configuration of a surface of a road on which the
system-equipped vehicle is moving. The quantity-of-vehicle state
sensor 37 works to measure the quantity of state of the
system-equipped vehicle. For instance, the quantity-of-vehicle
state sensor 37 measures the speed, the acceleration, or the yaw
rate of the system-equipped vehicle.
The navigation system 39 determines the location of the
system-equipped vehicle using a GPS. The map information storage 41
stores map information therein. The map information includes
information about gradients of given locations on the map. The
driver-assistance system 43 works to perform a known
driver-assistance operation, such as a lane-keeping assist
operation, using a road parameter, as calculated by the road
parameter calculator 1. The road parameter, as used in this
embodiment, is a parameter representing the configuration of a
road, such as a straight or a curved road, on which the
system-equipped vehicle is positioned.
2 Operation Executed by Road Parameter Calculator
Operations cyclically executed at a given interval by the road
parameter calculator 1 will be described below with reference to
FIGS. 3 to 6.
FIGS. 3(a) and 3(b) represent a flowchart of a sequence of logical
steps or a road parameter calculating program performed by the road
parameter calculator 1. After entering the program, the routine
proceeds to step S1 wherein the image acquiring unit 7 acquires an
image of a forward view of the system-equipped vehicle using the
camera 33.
The routine proceeds to step S2 wherein the edge-point extracting
unit 9 works to detect or extract edge points from the image, as
acquired in step 1. Each of the edge points, as referred to herein,
is expressed by a dot or pixel whose difference in brightness level
between itself and an adjacent pixel on the image is greater than a
given level.
The routine proceeds to step S3 wherein from the edge points
extracted in step 2, ones which have a higher probability that they
arise from a lane line (which will also be referred to as a vehicle
lane line) defining a lane on a road in which the system-equipped
vehicle is now traveling is selected.
Specifically, the operation in step S3 is achieved in the following
way. A Hough transform is performed on the edge points derived in
step S2 to determine lane line candidates. From the lane line
candidates, one which have a high probability that they represent
the vehicle lane line are selected using positions and directions
of the lane line candidates relative to the system-equipped
vehicle. Ones of the edge points which correspond to the selected
lane line candidate are derived.
The routine then proceeds to step S4 wherein the gradient detecting
unit 13 works to determine a change in inclination or gradient of
the road in front of the system-equipped vehicle (which will also
be referred to below as a gradient change). The operation in step
S4 will be described in detail with reference to FIG. 4. The
gradient, as referred to in this disclosure, is a gradient of the
road in a direction in which the system-equipped vehicle is moving
forward. The gradient usually includes an ascending gradient or a
descending gradient. The fact that there is the gradient change
means that the gradient of the road in front of the system-equipped
vehicle is different from that where the system-equipped vehicle
now exists.
For example, the gradient change occurs in a case where the
system-equipped vehicle now exists on a flat or horizontal surface
of the road, and there is an uphill or a downhill in front the
system-equipped vehicle. Alternatively, the gradient change may
occur in a case where the system-equipped vehicle is currently
located on an upward slope, and there is an upward slope with a
greater gradient, a horizontal road, or a downward slope in front
of the system-equipped vehicle. The gradient change may also occur
in a case where the system-equipped vehicle is currently located on
a downward slope, and there is a downward slope with a greater
gradient, a horizontal road, or an upward slope in front of the
system-equipped vehicle.
The gradient detecting unit 13 obtains the image 45, as illustrated
in FIG. 4, using the camera 33. The gradient detecting unit 13
calculates the first vanishing point 47 on the image 45. The first
vanishing point 47 is an intersection of the right and left vehicle
lane lines 49 which are lines on the image 45 defining a lane on
which the system-equipped vehicle is positioned.
The memory 5 stores the second vanishing point 51 in advance. The
second vanishing point 51 is defined as being a vanishing point
(i.e., an intersection of the right and left vehicle lane lines 49)
when the system-equipped vehicle is moving on a flat and horizontal
road surface. If the system-equipped vehicle travels at a long
distance, the second vanishing point 51 may be updated in a
learning mode.
The gradient detecting unit 13 continuously determines a positional
relation between the first vanishing point 47 and the second
vanishing point 51 in a vertical direction on the image 45.
The routine proceeds to step S5 wherein the gradient detecting unit
13 analyzes the vertical positional relation between the first
vanishing point 47 and the second vanishing point 51, as derived in
step S4, to determine whether there is the gradient change or
not.
Specifically, if the first vanishing point 47 remains above or
below the second vanishing point 51, it is determined that there is
the gradient change. Alternatively, if the first vanishing point 47
coincides with the second vanishing point 51, it is determined that
there is no gradient change. If the first vanishing point 47 moves
above or below the second vanishing point 51 cyclically at a short
interval, it is also determined that there is not gradient change.
This is thought of as arising from pitching motion of the
system-equipped vehicle. If a YES answer is obtained in step S5
meaning that there is the gradient change, then the routine
proceeds to step S6. Alternatively, if a NO answer is obtained in
step S5, then the routine proceeds to step S10.
In step S6, the modeling unit 15 produces a first model that is a
model for use in calculating the road parameter using the edge
points through a Kalman filter in step S17 which will be described
later in detail. The first model is a model which is defined by an
algorithm used in the Kalman filter and represents the
configuration of a road. The first model is designed in terms of a
road which extends more straight than the second model. This
modeling may be achieved by decreasing the degree or order in a
polynomial used in the Kalman filter in a way, as taught in
US2016/0148059 A1, filed on Nov. 23, 2015, assigned to the same
assignee as that of this application, disclosure of which is
totally incorporated herein by reference.
The routine proceeds to step S7 wherein the response setting unit
17 works to set the response rate or responsiveness of the Kalman
filter to be lower than that determined in step S11 which will be
described later in detail. The responsiveness of the Kalman filter
is a speed or rate of response of the Kalman filter to input of the
edge points and used in calculating the road parameter in the
following step S17. The lower the responsiveness of the Kalman
filter, the greater an effect of the road parameter, as derived
previously, on the road parameter, as calculated currently, thereby
resulting in a decrease in change in the road parameter.
The routine proceeds to step S8 wherein the edge-point limiting
unit 19 works to delimit a range of the edge points used in the
following step S17. This operation will be described below using
FIG. 5. FIG. 5 is a bird eye's view. "L1" and "L2" represent
locations of lines defined in front of the system-equipped vehicle
53 in the traveling direction F thereof. The line L1 is located
closer to the system-equipped vehicle 53 than the line L2 is.
The edge-point limiting unit 19 uses in the following step S17 ones
of the edge points selected in step S3 which lie closer to the
system-equipped vehicle than the line L1 does.
If a NO answer is obtained in step S5 meaning that there is no
gradient change, so that the operation in step S8 is not executed,
ones of the edge points selected in step S3 which lie closer to the
system-equipped vehicle 53 than the line L2 does are used in the
following step S17.
As apparent from the above discussion, when detecting the gradient
change, the edge-point limiting unit 19 defines the range in which
the edge points should be used in the following step S17 to be
closer to the system-equipped vehicle 53 than when there is no
gradient change.
Referring back to FIG. 3(a), after step S8, the routine proceeds to
step S9 wherein the threshold determining unit 29 prepares the
first threshold value. The first threshold value is used in the
following step S15. The first threshold value is smaller than a
second threshold value which will be described later in detail.
If a NO answer is obtained in step S5 meaning that there is no
gradient change, then the routine proceeds to step S10 wherein the
modeling unit 15 defines the second model that is a mode for use in
calculating the road parameter using the edge points through the
Kalman filter in the following step S17.
The routine proceeds to step S11 wherein the response setting unit
17 sets the value of the responsiveness (i.e., a response rate) of
the Kalman filter to a normal value. The normal value is selected
to be a higher response rate than that determined in step S7.
The routine proceeds to step S12 wherein the threshold determining
unit 29 prepares the second threshold value. The second threshold
value is used in the following step S15. The second threshold value
is smaller than the first threshold value.
The routine proceeds to step S13 wherein the lane line producing
unit 21 works to produce the vehicle lane line using the edge
points. The edge points, as selected in step S3, are used in step
S13. When the operation in step S8 is executed, ones of the edge
points selected in step S3 which lie closer to the system-equipped
vehicle than the line L1 does are used in step S13.
The routine proceeds to step S14 wherein the likelihood calculating
unit 23 calculates a likelihood that the vehicle lane line, as
derived in step S13, is a branch line (i.e., a lane line of a
branch road). The likelihood is calculated using at least one of a
curvature of the vehicle lane line, a yaw angle, and an offset. The
yaw angle is an angle which the traveling direction F of the
system-equipped vehicle 53 illustrated in FIG. 6 makes with a
longitudinal direction of the vehicle lane line 49. FIG. 6 is a
bird's eye view. The offset is a distance between the
system-equipped vehicle 53 and the vehicle lane line 49 (i.e., the
right lane line in the example of FIG. 6) in the widthwise
direction of the road.
Referring back to FIG. 3(b), after step S14, the routine proceeds
to step S15 wherein the branch line determining unit 25 determines
whether the likelihood, as calculated in step S14, is greater than
a threshold value or not. The threshold value used in step S15 is
the first threshold value when the operation in step S9 has been
executed or the second threshold value when the operation in step
S12 has been executed.
If a YES answer is obtained in step S15 meaning that the likelihood
is greater than the threshold value, in other words, the vehicle
lane line, as calculated in step S13, is the branch line, then the
routine proceeds to step S16. Alternatively, if a NO answer is
obtained meaning that the likelihood is less than or equal to the
threshold value, in other words, the vehicle lane line, as
calculated in step S13, is not the branch line, then the routine
proceeds to step S17.
In step S16, the edge-point removing unit 27 excludes the edge
points on the vehicle lane line, as determined as the branch line
in step S15, from the edge points for use in the following step
S17.
In step S17, the road parameter calculating unit 11 calculates the
road parameter using the edge points through the Kalman filter. In
other words, the edge points are inputted to the Kalman filter to
derive the road parameter representing the configuration of the
road. Basically, the edge points, as used in step S17, are the edge
points selected in step S3. However, when the operation in step S8
has been executed, ones of the edge points selected in step S3
which lie in a range located closer to the system-equipped vehicle
than the line L1 is are selected. Alternatively, when the operation
in step S16 has been executed, from the edge points, ones which lie
on the branch line are removed.
The model used in the Kalman filter is set to the first model when
the operation in step S6 has been executed or the second model when
the operation in step S10 has been executed.
The responsiveness or response rate of the Kalman filter is
selected to be low when the operation in step S7 has been executed
or to be normal when the operation in step S11 has been
executed.
After step S17, the routine proceeds to step S18 wherein the output
unit 31 outputs the road parameter, as calculated in step S17, to
the driver-assistance system 43.
3 Effects of the Road Parameter Calculator
1A When the gradient change is detected, the road parameter
calculator 1 uses the first model which is designed to model a road
which is more straight, in other words, has a reduced curvature
than when the gradient change is not detected. This minimizes the
effect of the gradient change on calculation of the road parameter.
1B When the gradient change is detected, the road parameter
calculator 1 changes the response rate of the Kalman filter to be
slower than when the gradient change is not detected. This
minimizes the effect of the gradient change on the calculation of
the road parameter. 1C When there is the gradient change, the
calculation of the road parameter using the edge points located far
away from the system-equipped vehicle usually result in an increase
in effect of the gradient change on the value of the road
parameter. The road parameter calculator 1, therefore, works to
delimit the range of the edge points for use in calculating the
road parameter to be closer to the system-equipped vehicle when the
gradient change is detected than when no gradient change is
detected. This minimizes the adverse effect of the gradient change
on the calculation of the road parameter. 1D The road parameter
calculator 1 is also designed to change the threshold value for use
in determining whether the vehicle lane line is the branch line or
not when the gradient change is detected to be greater than when no
gradient change is detected. This minimizes an error in determining
that the vehicle lane line is the branch line when there is the
gradient change, thereby ensuring the stability in calculating the
road parameter correctly regardless of the gradient change. 1E The
road parameter calculator 1 detects the gradient change using the
vertical positional relation between the first vanishing point 47
and the second vanishing point 51, thereby facilitating the
detection of the gradient change and enhancing the accuracy of such
detection.
Modifications
While the present invention has been disclosed in terms of the
preferred embodiment in order to facilitate better understanding
thereof, it should be appreciated that the invention can be
embodied in various ways without departing from the principle of
the invention. Therefore, the invention should be understood to
include all possible embodiments and modifications to the shown
embodiment which can be embodied without departing from the
principle of the invention as set forth in the appended claims.
(1) In step S6, when two vehicle lane lines: the right and left
vehicle lane lines are being detected, a first model A may be
selected. Alternatively, when only one of the vehicle lane lines is
being detected, a first model B may be selected. The first model B
is designed to model a road forward extending more straight, in
other words, a road with a reduced curvature than the first model
A. The first model A is designed to model a road extending more
straight than the second model.
The selection of the model in the above way greatly decreases the
effect of the gradient change on the calculation of the road
parameter when there is the gradient change, and only one of the
vehicle lane lines is being detected.
(2) The gradient detecting unit 13 also determines the degree of
the gradient change as well as detection of the gradient change.
Specifically, the gradient detecting unit 13 may also work as a
degree-of-gradient calculating unit. In step S7, the response rate
may be lowered with an increase in degree of the gradient change.
This also enables the response rate of the Kalman filter as a
function of the degree of the gradient change. (3) In step S8, the
edge points may be limited in the following way. FIG. 10 is a bird
eye's view. In case of the left vehicle lane line 49, the edge
points are detected to the location (i.e., a broken line) M1. In
case of the right vehicle lane line 49, the edge points are
detected to the location (i.e., a broken line) M2. The lines M1 and
M2 are located in front of the system-equipped vehicle 53 in the
traveling direction F. The line M1 is closer to the system-equipped
vehicle 53 than the line M2 is.
The edge-point limiting unit 19 limits a right and a left range in
which the edge points on the right and left vehicle lane lines 49
are used in step S17 to be closer to the system-equipped vehicle
than the line M1 is. In other words, the edge-point limiting unit
19 sets the right and left ranges of the edge points on the right
and left sides of the system-equipped vehicle for use in step S17
to be equal to each other.
The use of the edge points lying between the lines M1 and M2 on one
of the right and left sides of the system-equipped vehicle results
in a greater effect of the gradient change on the calculation of
the road parameter. This problem is, therefore, alleviated by
defining the ranges of the edge points on the right and left sides
of the system-equipped vehicle for use in step S17 to be closer to
the system-equipped vehicle than the line M1 is.
(4) In steps S4 and S5, the determination of whether there is the
gradient change or not may be achieved in another way. For
instance, the gradient change may be detected using a change in
positional relation between a vehicle traveling ahead of the
system-equipped vehicle and the vehicle lane line (either of the
right or left vehicle lane line) in the width-wise direction of the
road.
The vehicle lane lines 49 in FIG. 7 may be sometimes viewed on the
image as branch lines. FIG. 7 is a bird eye's view into which an
image captured by the camera 33 is converted. The distance between
the preceding vehicle 55 and one of the vehicle lane lines 49 is
expressed by d1. The distance between the preceding vehicle 55 and
the other vehicle lane line 49 is expressed by d2. The sum of the
distances d1 and d2 is expressed by ds. If the vehicle lane lines
49 are actually the branch lines, it will cause the distance ds to
change with time. Alternatively, if the vehicle lane lines 49 are
not actually the branch lines, in other words, look like as if they
are the branch lines due to the gradient change, it will cause the
distance ds to be kept constant.
Using the above conditions, when the vehicle lane lines 49 which
look like the branch lines appear on the image, and the distance ds
is kept constant with time, the gradient detecting unit 13 may
determine that there is the gradient change. This method
facilitates the detection of the gradient change with an increased
accuracy.
(5) In steps 4 and 5, the determination of whether there is the
gradient change or not may be achieved by using a positional
relation between a locus or trace of a preceding vehicle and the
vehicle lane line.
The vehicle lane lines 49 in FIG. 8 may be sometimes viewed on the
image as branch lines. FIG. 8 is a bird eye's view into which an
image captured by the camera 33 is converted. The trace of the
preceding vehicle 55 traveling ahead of the system-equipped vehicle
is expressed by a broken line 57. When the vehicle lane lines
actually represent the branch lines, the trace 57 will extend along
either one of the vehicle lane lines 49. Alternatively, when the
vehicle lane lines are not actually the branch lines, in other
words, look like as if they are the branch lines due to the
gradient change, it will cause the trace 57 to lie at the middle
between the right and left vehicle lane lines 49.
Using the above conditions, when the vehicle lane lines 49 which
look like the branch lines appear on the image, and the trace 57
lies at the middle between the right and left vehicle lane lines
49, the gradient detecting unit 13 may determine that there is the
gradient change. This method facilitates the detection of the
gradient change with an increased accuracy.
(6) In steps S4 and S5, the determination of whether there is the
gradient change or not may be achieved in another way. For
instance, the gradient change may be detected using a combination
of curvatures of the right and left vehicle lane lines.
Specifically, when there is the gradient change, the apparent
curvatures (i.e., orientation of curves) of the right and left
vehicle lane lines 49 on the image are, as can be seen in FIG. 9,
usually opposite each other. FIG. 9 is a bird eye's view into which
an image captured by the camera 33 is converted. In the example of
FIG. 9, the right vehicle lane line 49 is curved in the right
direction, while the left vehicle lane line 49 is curved in the
left direction.
Based on the above fact, when the apparent curvatures of the right
and left vehicle lane lines 49 on the image are opposite each
other, the gradient detecting unit 13 determines that there is the
gradient change. This method facilitates the detection of the
gradient change with an increased accuracy.
(7) The determination of whether there is the gradient change or
not in steps S4 and S5 may also be achieved in another way. For
instance, the gradient change may be detected using information
about gradients of roads recorded in a map. Specifically, the
gradient detecting unit 13 may acquire the current location of the
system-equipped vehicle using the navigation system 39, read
gradients at the location of the system-equipped vehicle and at a
given location in front of the system-equipped vehicle out of the
map information storage 41, and then compare the gradients read out
of the map information storage 41 to determine whether there is the
gradient change or not. This method also facilitates the detection
of the gradient change with an increased accuracy. (8) The
determination of whether there is the gradient change or not in
steps S4 and S5 may also be achieved in another way. For instance,
the gradient detecting unit 13 may obtain the configuration of a
surface of a road in front of the system-equipped vehicle using the
surroundings sensor 35 and analyze it to determine whether there is
the gradient change or not. This method also facilitates the
detection of the gradient change with increased accuracy. (9) In
the above embodiment, a plurality of functions of one of the
components of the road parameter calculator 1 may be shared with
two or more of the components. A single function of one of the
components may be achieved by two or more of the other components.
Alternatively, two or more functions of two of more of the
components may be performed by only one of the components. A single
function performed by two or more of the components may be achieved
by one of the components. The components of the above embodiment
may be partially omitted. (10) The above described road parameter
calculator 1 may be achieved in one of various modes: a system
equipped with the road parameter calculator 1, a logical program
executed by a computer which realizes the road parameter calculator
1, a non-transitory tangible storage medium, such as a
semiconductor memory, which stores the program, and a road
parameter calculating method.
* * * * *